##################DESCRIPTIVE STATISTICS
sum vote_counted_fairly_reverse 
sum electoral_responsiveness_reverse
sum efficacy_dontcare
sum efficacy_nosay
sum democratic_ideals_reverse
sum age
sum gender_male
sum education
sum income
sum ideology
sum race_white
sum hispanic
sum news_consumption
sum political_interest
sum polknowledge_scale
sum swingstate
tab vote_choice
tab partyID_three


##################TABLE 2:ANALYSIS OF MAIN RESULTS
Model 1: VOTE COUNTED FAIRLY
ologit vote_counted_fairly_reverse fraud_treatment b3.partyID_three b1.vote_choice age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale swingstate if voter_status==1
est sto m1

Model 2: ELECTORAL RESPONSIVENESS
ologit electoral_responsiveness_reverse fraud_treatment b3.partyID_three b1.vote_choice age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m2

Model 3: EFFICACY DONTCARE
ologit efficacy_dontcare fraud_treatment b3.partyID_three b1.vote_choice age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m3

Model 4: EFFICACY NOSAY
ologit efficacy_nosay fraud_treatment b3.partyID_three b1.vote_choice age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m4

Model 5: DEMOCRATIC IDEALS
ologit democratic_ideals_reverse fraud_treatment b3.partyID_three b1.vote_choice age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m5


##################FIGURE 1:COEFFICIENT PLOT FOR TREATMENT EFFECTS
ologit vote_counted_fairly_reverse fraud_treatment b3.partyID_three b1.vote_choice age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale swingstate if voter_status==1
est sto m1

ologit electoral_responsiveness_reverse fraud_treatment b3.partyID_three b1.vote_choice age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m2

ologit efficacy_dontcare fraud_treatment b3.partyID_three b1.vote_choice age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m3

ologit efficacy_nosay fraud_treatment b3.partyID_three b1.vote_choice age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m4

ologit democratic_ideals_reverse fraud_treatment b3.partyID_three b1.vote_choice age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m5

coefplot (m1,label("DV: Belief that Voting was Fair")) (m2,label("DV: Belief that Elections Make Government Pay Attention to People")) (m3,label("DV: Level of Political Efficacy (Item: Public Officials Don't Care About People)")) (m4,label("DV: Level of Political Efficacy (Item: People Have No Say About Government)")) (m5,label("DV: Level of Support for Democracy")), keep(fraud_treatment) xline(0) ylabel("") level(95) title("Figure 1: Confounder-Adjusted Treatment Effect Sizes") xscale(range(-.70 0.15)) legend(col(1))


##################FIGURE 2:ESTIMATED PROBABILITY CHANGES
ologit electoral_responsiveness_reverse fraud_treatment b3.partyID_three b1.vote_choice age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale swingstate
margins, at(fraud_treatment=(0 1)) predict(outcome(3))
marginsplot, title (Panel A: Belief that Elections Make Gov. Pay Attention, size(medium)) subtitle(" ") xtitle("") xscale(range(-.5 1.5)) xlabel(1 "Treatment" 0 "Control") ytitle(Probability of DV = Above Midpoint) plotopts(connect(none)) saving(panel-a.gph)

ologit efficacy_dontcare fraud_treatment b3.partyID_three b1.vote_choice age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale swingstate
margins, at(fraud_treatment=(0 1)) expression(predict(outcome(4)) + predict(outcome(5)))
marginsplot, title (Panel B: Level of Efficacy (Item: Officials Don't Care), size(medium)) subtitle(" ") xtitle("") xscale(range(-.5 1.5)) xlabel(1 "Treatment" 0 "Control") ytitle(Probability of DV = Above Midpoint) plotopts(connect(none)) saving(panel-b.gph)

ologit efficacy_nosay fraud_treatment b3.partyID_three b1.vote_choice age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale swingstate
margins, at(fraud_treatment=(0 1)) expression(predict(outcome(4)) + predict(outcome(5)))
marginsplot, title (Panel C: Level of Efficacy (Item: People Have No Say), size(medium)) subtitle(" ") xtitle("") xscale(range(-.5 1.5)) xlabel(1 "Treatment" 0 "Control") ytitle(Probability of DV = Above Midpoint) plotopts(connect(none)) saving(panel-c.gph)

ologit democratic_ideals_reverse fraud_treatment b3.partyID_three b1.vote_choice age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale swingstate
margins, at(fraud_treatment=(0 1)) expression(predict(outcome(3)) + predict(outcome(4)))
marginsplot, title (Panel D: Level of Support for Democracy, size(medium)) subtitle(" ") xtitle("") xscale(range(-.5 1.5)) xlabel(1 "Treatment" 0 "Control") ytitle(Probability of DV = Above Midpoint) plotopts(connect(none)) saving(panel-d.gph)

graph combine panel-a.gph panel-b.gph panel-c.gph panel-d.gph, xcommon title(Figure 2: Effect of Treatment on Democratic Attitudes) iscale(.62)







##################TABLE 3:ANALYSIS OF INTERACTIVE RELATIONSHIPS (PARTY ID)
Model 6: VOTE COUNTED FAIRLY
ologit vote_counted_fairly_reverse fraud_treatment##b3.partyID_three b1.vote_choice age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale swingstate if voter_status==1
est sto m6
lrtest m1 m6
margins, dydx(fraud_treatment) at(partyID_three=(1(1)3)) expression(predict(outcome(4)) + predict(outcome(5)))

Model 7: ELECTORAL RESPONSIVENESS
ologit electoral_responsiveness_reverse fraud_treatment##b3.partyID_three b1.vote_choice age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m7
lrtest m2 m7

Model 8: EFFICACY DONTCARE
ologit efficacy_dontcare fraud_treatment##b3.partyID_three b1.vote_choice age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m8
lrtest m3 m8

Model 9: EFFICACY NOSAY
ologit efficacy_nosay fraud_treatment##b3.partyID_three b1.vote_choice age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m9
lrtest m4 m9

Model 10: DEMOCRATIC IDEALS
ologit democratic_ideals_reverse fraud_treatment##b3.partyID_three b1.vote_choice age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m10
lrtest m5 m10



##################TABLE 4:ANALYSIS OF INTERACTIVE RELATIONSHIPS (VOTE CHOICE)
Model 11: VOTE COUNTED FAIRLY
ologit vote_counted_fairly_reverse fraud_treatment##b1.vote_choice b3.partyID_three age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale swingstate if voter_status==1
est sto m11
lrtest m1 m11
margins, dydx(fraud_treatment) at(vote_choice=(1(1)3)) expression(predict(outcome(4)) + predict(outcome(5)))

Model 12: ELECTORAL RESPONSIVENESS
ologit electoral_responsiveness_reverse fraud_treatment##b1.vote_choice b3.partyID_three age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m12
lrtest m2 m12

Model 13: EFFICACY DONTCARE
ologit efficacy_dontcare fraud_treatment##b1.vote_choice b3.partyID_three age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m13
lrtest m3 m13

Model 14: EFFICACY NOSAY
ologit efficacy_nosay fraud_treatment##b1.vote_choice b3.partyID_three age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m14
lrtest m4 m14

Model 15: DEMOCRATIC IDEALS
ologit democratic_ideals_reverse fraud_treatment##b1.vote_choice b3.partyID_three age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m15
lrtest m5 m15




##################TABLE 5:ANALYSIS OF INTERACTIVE RELATIONSHIPS (IDEOLOGY)
Model 16: VOTE COUNTED FAIRLY
ologit vote_counted_fairly_reverse fraud_treatment##c.ideology b3.partyID_three b1.vote_choice age gender_male education income race_white hispanic news_consumption political_interest polknowledge_scale swingstate if voter_status==1
est sto m16
lrtest m1 m16

Model 17: ELECTORAL RESPONSIVENESS
ologit electoral_responsiveness_reverse fraud_treatment##c.ideology b3.partyID_three b1.vote_choice age gender_male education income race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m17
lrtest m2 m17

Model 18: EFFICACY DONTCARE
ologit efficacy_dontcare fraud_treatment##c.ideology b3.partyID_three b1.vote_choice age gender_male education income race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m18
lrtest m3 m18

Model 19: EFFICACY NOSAY
ologit efficacy_nosay fraud_treatment##c.ideology b3.partyID_three b1.vote_choice age gender_male education income race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m19
lrtest m4 m19

Model 20: DEMOCRATIC IDEALS
ologit democratic_ideals_reverse fraud_treatment##c.ideology b3.partyID_three b1.vote_choice age gender_male education income race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m20
lrtest m5 m20




##################TABLE 6:ANALYSIS OF INTERACTIVE RELATIONSHIPS (SWINGSTATE)
Model 21: VOTE COUNTED FAIRLY
ologit vote_counted_fairly_reverse fraud_treatment##swingstate b3.partyID_three b1.vote_choice age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale if voter_status==1
est sto m21
lrtest m1 m21

Model 22: ELECTORAL RESPONSIVENESS
ologit electoral_responsiveness_reverse fraud_treatment##swingstate b3.partyID_three b1.vote_choice age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale
est sto m22
lrtest m2 m22

Model 23: EFFICACY DONTCARE
ologit efficacy_dontcare fraud_treatment##swingstate b3.partyID_three b1.vote_choice age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale
est sto 23
lrtest m3 m23

Model 24: EFFICACY NOSAY
ologit efficacy_nosay fraud_treatment##swingstate b3.partyID_three b1.vote_choice age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale
est sto 24
lrtest m4 m24

Model 25: DEMOCRATIC IDEALS
ologit democratic_ideals_reverse fraud_treatment##swingstate b3.partyID_three b1.vote_choice age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale
est sto m25
lrtest m5 m25




##################FIGURE 3:INTERACTIVE RELATIONSHIP (PARTY ID; VOTE COUNTED FAIRLY REVERSE)
ologit vote_counted_fairly_reverse fraud_treatment##b3.partyID_three b1.vote_choice age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale swingstate if voter_status==1
margins, dydx(fraud_treatment) at(partyID_three=(1(1)3)) expression(predict(outcome(4)) + predict(outcome(5)))
marginsplot, title (Panel A: Treatment Effects by Party ID)  xtitle("Party ID") xscale(range(.5 3.5)) xlabel(1 "Democrat" 2 "Independent" 3 "Republican") ytitle(Estimated Treatment Effect) plotopts(connect(none)) yline(0)  saving(f3.panel-a.gph)

ologit vote_counted_fairly_reverse fraud_treatment##b1.vote_choice b3.partyID_three  age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale swingstate if voter_status==1
margins, dydx(fraud_treatment) at(vote_choice=(1(1)3)) expression(predict(outcome(4)) + predict(outcome(5)))
marginsplot, title (Panel B: Treatment Effects by Vote Choice)  xtitle("Candidate Preference") xscale(range(.5 3.5)) xlabel(1 "Trump" 2 "Biden" 3 "Other") ytitle(Estimated Treatment Effect) plotopts(connect(none)) yline(0)  saving(f3.panel-b.gph)

graph combine f3.panel-a.gph f3.panel-b.gph, xcommon title(Figure 3: Treatment Effect Differences (DV: Belief that Voting was Fair)) iscale(.65)




##################FIGURE 4:INTERACTIVE RELATIONSHIP (VOTE COUNTED FAIRLY REVERSE)
Non-Trump Republicans
ttest vote_counted_fairly_binary if partyID_three==3 & vote_choice!=1, by(fraud_treatment)
Non-Trump Non-Republicans
ttest vote_counted_fairly_binary if partyID_three!=3 & vote_choice!=1, by(fraud_treatment)
Trump Non-Republicans
ttest vote_counted_fairly_binary if partyID_three!=3 & vote_choice==1, by(fraud_treatment)
Trump Republicans
ttest vote_counted_fairly_binary if partyID_three==3 & vote_choice==1, by(fraud_treatment)
INTERESTING FINDING WHICH IS NOT USED: ologit vote_counted_fairly_reverse fraud_treatment##c.news_consumption age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale swingstate if voter_status==1 & voter_group==4
INTERESTING FINDING WHICH IS NOT USED: ologit vote_counted_fairly_reverse fraud_treatment##c.political_interest age gender_male education income ideology race_white hispanic news_consumption political_interest polknowledge_scale swingstate if voter_status==1 & voter_group==4

USE DATA FOR FIGURE 4:
serrbar proportion std_error voter_group_numeric, scale(1.96) yline(0) xscale(range(0.5 4.5)) xtitle("Voter Group") ylabel(,labsize(small)) ytitle("Change in Proportion") xlabel(1 "Non-Trump GOP" 2 "Non-Trump Non-GOP" 3"Trump Non-GOP" 4 "Trump GOP",  labsize(small)) title("Figure 4: Effect of Treatment on Belief that Voting was Fair") subtitle("By Nested Voter Group")





##################TABLE R1:ANALYSIS OF INTERACTIVE RELATIONSHIPS (PARTY ID)
Model 6: VOTE COUNTED FAIRLY
ologit vote_counted_fairly_reverse fraud_treatment##b3.partyID_three age gender_male education income race_white hispanic news_consumption political_interest polknowledge_scale swingstate if voter_status==1
est sto m6
lrtest m1 m6
margins, dydx(fraud_treatment) at(partyID_three=(1(1)3)) expression(predict(outcome(4)) + predict(outcome(5)))

Model 7: ELECTORAL RESPONSIVENESS
ologit electoral_responsiveness_reverse fraud_treatment##b3.partyID_three age gender_male education income race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m7
lrtest m2 m7

Model 8: EFFICACY DONTCARE
ologit efficacy_dontcare fraud_treatment##b3.partyID_three age gender_male education income race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m8
lrtest m3 m8

Model 9: EFFICACY NOSAY
ologit efficacy_nosay fraud_treatment##b3.partyID_three age gender_male education income race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m9
lrtest m4 m9

Model 10: DEMOCRATIC IDEALS
ologit democratic_ideals_reverse fraud_treatment##b3.partyID_three age gender_male education income race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m10
lrtest m5 m10



##################TABLE R2:ANALYSIS OF INTERACTIVE RELATIONSHIPS (VOTE CHOICE)
Model 11: VOTE COUNTED FAIRLY
ologit vote_counted_fairly_reverse fraud_treatment##b1.vote_choice age gender_male education income race_white hispanic news_consumption political_interest polknowledge_scale swingstate if voter_status==1
est sto m11
lrtest m1 m11
margins, dydx(fraud_treatment) at(vote_choice=(1(1)3)) expression(predict(outcome(4)) + predict(outcome(5)))

Model 12: ELECTORAL RESPONSIVENESS
ologit electoral_responsiveness_reverse fraud_treatment##b1.vote_choice age gender_male education income race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m12
lrtest m2 m12

Model 13: EFFICACY DONTCARE
ologit efficacy_dontcare fraud_treatment##b1.vote_choice age gender_male education income race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m13
lrtest m3 m13

Model 14: EFFICACY NOSAY
ologit efficacy_nosay fraud_treatment##b1.vote_choice age gender_male education income race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m14
lrtest m4 m14

Model 15: DEMOCRATIC IDEALS
ologit democratic_ideals_reverse fraud_treatment##b1.vote_choice age gender_male education income race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m15
lrtest m5 m15




##################TABLE R3:ANALYSIS OF INTERACTIVE RELATIONSHIPS (IDEOLOGY)
Model 16: VOTE COUNTED FAIRLY
ologit vote_counted_fairly_reverse fraud_treatment##c.ideology age gender_male education income race_white hispanic news_consumption political_interest polknowledge_scale swingstate if voter_status==1
est sto m16
lrtest m1 m16

Model 17: ELECTORAL RESPONSIVENESS
ologit electoral_responsiveness_reverse fraud_treatment##c.ideology age gender_male education income race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m17
lrtest m2 m17

Model 18: EFFICACY DONTCARE
ologit efficacy_dontcare fraud_treatment##c.ideology age gender_male education income race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m18
lrtest m3 m18

Model 19: EFFICACY NOSAY
ologit efficacy_nosay fraud_treatment##c.ideology age gender_male education income race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m19
lrtest m4 m19

Model 20: DEMOCRATIC IDEALS
ologit democratic_ideals_reverse fraud_treatment##c.ideology age gender_male education income race_white hispanic news_consumption political_interest polknowledge_scale swingstate
est sto m20
lrtest m5 m20




